Pyroptosis-related pattern depicting tumor microenvironment and association with prognosis and efficacy in patients with hepatocellular carcinoma receiving combined molecular targeted therapy and immune checkpoint inhibitors: A multi-omics study.

Abstract

e16194 Background: Molecular targeted therapy combined with immunotherapy is among the recommended treatment strategies for initially unresectable hepatocellular carcinoma (HCC). Pyroptosis is believed to be closely associated with immune cell infiltration in many tumors. We attempted to utilize a readily testable set of pyroptosis-related genes (PRGs) to develop genetic features predicting the efficacy of targeted combined immunotherapy in HCC, aiming to identify individuals sensitive to treatment. Methods: We conducted bioinformatics analysis on 53 PRGs using TCGA-LIHC and GTEx databases to construct PIs (Pyroptosis-Immune score). External and internal validation were performed using the GEO database and sequencing data from our center. Paired single-cell data were used to validate the predictive efficacy of PIs on the effectiveness of immune treatment. A matched cohort of 32 vs 32 individuals was established based on whether they underwent surgery after targeted combined immunotherapy. Immunohistochemistry staining analysis was conducted on pre-treatment biopsy samples, and protein expression was converted to corresponding gene expression levels. P-PIs (Pathological PIs) were calculated to verify its clinical applicability. The surgical group was assessed based on postoperative pathological efficacy, and the non-surgical group was evaluated for progression-free survival (PFS) through imaging. Results: We identified 6 PRGs closely associated with prognosis, immune cell infiltration, and the tumor microenvironment, constructing PIs. Log-Rank risk regression analysis suggested worse prognosis in the high PIs group (OS, HR = 3.44, 95% CI: 2.33-5.05, p < 0.01; PFS, HR = 2.00, 95% CI: 1.44-2.81, p < 0.01). GSE76427 dataset (n = 80; OS: HR = 1.61, 95% CI: 1.15-2.26, P < 0.01; PFS: HR = 1.44, 95% CI: 1.03-2.02, P = 0.03) and our center's sequencing data (n = 43; OS: HR = 1.53, 95% CI: 1.10-2.14, P = 0.01; PFS: HR = 1.55, 95% CI: 1.08-2.21, P = 0.02) indicated that the PIs could predict the prognosis for HCC. Single-cell sequencing data showed that tumor cells with high PI scores inhibited the function of CD8+ T cells (R = -0.33), while MDSC cells (R = 0.22), and M2 macrophage functions (R = 0.22) were enhanced. The P-PIs of operation group was higher (P = 0.01). The P-PIs in the surgical group was positively correlated with the degree of pathological remission (R = 0.95, P < 0.01). The median PFS in the non-surgical group was 8.5 months, and the P-PIs AUC was 0.76 (95% CI: 0.66-0.83). Conclusions: We developed a PIs to predict efficacy of targeted combined immunotherapy in HCC based on pre-treatment biopsy pathology. The P-PIs could potentially facilitate clinical trials aiming at personalized efficacy prediction for conversion or adjuvant treatment in HCC patients.